Chemical Bonding Technology: Direct Investigation of ... · of hydrocarbon encapsulants such as ethylene vinyl acetate (EVA), the need ... D. HYDROTHERMAL AGING OF EVA/GLASS INTERFACE

5101-284 DOE/JPL-1012-120Flat-Plate Solar Distribution Category UC-63bArray Project

Chemical Bonding Technology:Direct Investigation ofInterfacial Bonds

J.L. KoenigF.J. BoerioE.P. PlueddemannJ. MillerP.B. WillisE.F. Cuddihy

(NASa-CK-176657) CHEMICA1 B O N D I N G N86-21584TECHNOLOGY: .'DIEECT I N V E S T I G A T I O N Of -.INTEBFACIft l BONDS (Jet Propulsion Lab.),54 p HC A'04-/ttf J>01 CSCL 07C Unclas

v i G 3/2 3 03 8 6J

January 1986

Prepared forU.S. Department of EnergyThrough an Agreement withNational Aeronautics and Space AdministrationbyJet Propulsion LaboratoryCalifornia Institute of TechnologyPasadena, California

JPL Publication 86-6

5101-284 DOE/JPL-1012-120

Flat-Plate Solar Distribution Category UC-63bArray Project

Chemical Bonding Technology:Direct Investigation ofInterfacial Bonds

J.L. KoenigF.J. BoerioE.P. PlueddemannJ. MillerP.B. WillisE.F. Cuddihy

January 1986

Prepared for

U.S. Department of Energy

Through an Agreement withNational Aeronautics and Space Administration

byJet Propulsion LaboratoryCalifornia Institute of TechnologyPasadena, California

JPL Publication 86-6

Prepared by the Jet Propulsion Laboratory, California Institute of Technology,for the U.S. Department of Energy through an agreement with the NationalAeronautics and Space Administration.

The JPL Flat-Plate Solar Array Project is sponsored by the U.S. Department ofEnergy and is part of the National Photovoltaics Program to initiate a majoreffort toward the development of cost-competitive solar arrays.

This report was prepared as an account of work sponsored by an agency of theUnited States Government. Neither the United States Government nor anyagency thereof, nor any of their employees, makes any warranty, express orimplied, or assumes any legal liability or responsibility for the accuracy, com-pleteness, or usefulness of any information, apparatus, product, or processdisclosed, or represents that its use would not infringe privately owned rights.

Reference herein to any specific commercial product, process, or service by tradename, trademark, manufacturer, or otherwise, does not necessarily constitute orimply its endorsement, recommendation, or favoring by the United StatesGovernment or any agency thereof. The views and opinions of authors expressedherein do not necessarily state or reflect those of the United States Governmentor any agency thereof.

This document reports on work done under NASA Task RE-152, Amendment419, DOE/NASA IAA No. DE-A101-85CE89008.

ABSTRACT

This is the third Flat-Plate Solar Array (FSA) Project document reportingon chemical bonding technology for terrestrial photovoltaic (PV) modules. Theimpetus for this work originated in the late 1970s when PV modules employingsilicone encapsulation materials were undergoing delamination during outdoorexposure. At that time, manufacturers were not employing adhesion promotersand, hence, module interfaces in common with the silicone materials were onlyin physical contact and therefore easily prone to separation if, for example,water were to penetrate to the interfaces. Delamination with siliconematerials virtually vanished when adhesion promoters, recommended by siliconemanufacturers, were used.

With the decrease in use of silicone encapsulants, and the increase in useof hydrocarbon encapsulants such as ethylene vinyl acetate (EVA), the needdeveloped for adhesion promoters specifically developed for these new materials.The adhesion promoters being developed for EVA-type materials are based onorganosilanes, which generate primary chemical bonds at the interface, that is,chemical bonding. These adhesion promoters are commonly referred to as"primers."

The first report on this subject (Chemical Bonding Technology forTerrestrial Solar Cell Modules, by E.P. Plueddeman, JPL 5101-132, datedSeptember 1, 1979) described the chemistry of primers based on organosilanechemistry, and the second report (Chemical Bonding Technology for TerrestrialPhotovoltaic Modules, by D.R. Coulter, E.F. Cuddihy, and E.P. Plueddeman, JPLPublication 83-86, dated November 15, 1983) described chemical bondingtheories, and also included a listing of candidate primer and adhesive systemsbeing investigated for all of the various module interfaces. This reportdescribes the activities related to the direct investigation of chemicallybonded interfaces.

111

ACKNOWLEDGMENT

The authors thank the following companies for courteously supplying solarcells used for the adhesion studies discussed in this Project report: SOLECInternational, Hawthorne, California; Spectrolab, Inc., Sylmar, California; andTideland Signal, Houston, Texas.

IV

ABBREVIATIONS AND ACRONYMS

AES Auger electron spectroscopy

DRIFT diffuse reflectance infrared spectroscopy

EDAX energy dispersive x-ray analysis

EVA ethylene vinyl acetate

FSA Flat-Plate Solar Array

FTIR Fourier transform infrared spectroscopy

IPN Interpenetrating Polymer Network

OH organic alcohols

OR organic ethers

PV photovoltaic

RAIR reflection absorption infrared spectroscopy

XPS x-ray photoelectron spectra

CONTENTS

I. INTRODUCTION 1

II. CHEMICAL BONDING BETWEEN GLASS AND ETHYLENEVINYL ACETATE (EVA) 3

A. FOURIER TRANSFORM INFRARED (FTIR) SPECTROSCOPY OFSURFACES AND INTERFACES 6

1. Avantages of FTIR for Surface Studies 6

2. Sampling Techniques for FTIR Spectroscopy Studies . . 6

3. Data Processing Techniques 7

B. EXPERIMENTAL PROCEDURES (MATERIALS) '. 7

C. FTIR SPECTROSCOPY STUDIES OF UNAGEDEVA/GLASS INTERFACES 8

1. Glass 8

2. Coupling Agent on Glass Surface 11

3. Hydrothermal Stability of the Glass/CouplingAgent Interface 13

4. EVA 14

5. EVA/Glass Interface 14

D. HYDROTHERMAL AGING OF EVA/GLASS INTERFACE 17

1. Mechanical Testing 17

2. Spectroscopic Testing 25

III. CHEMICAL BONDING BETWEEN SOLAR CELLS AND EVA 29

A. ALUMINUM MIRRORS ' 29

B. ALUMINUM MIRRORS WITH EVA COATINGS 31

C. ALUMINIZED SOLAR CELLS 33

D. REFLECTION ABSORPTION INFRARED SPECTROSCOPY STUDIES .... 39

IV. REFERENCES 43

vii PRECEDING PAGE BLANK NOT RLMED

Figures

1. Bonding of Silane Coupling Agents to Glass 5

2. Interdiffusion Model for a Silane-PrimedEVA/Glass Joint 5

3. Diffuse Reflectance Spectrum of Material Componentsand Composites 10

4. Diffuse Reflectance FTIR Spectra of Y-MPS/AmineModified Particulate Glass, the Untreated ParticulateGlass, and the Difference Spectrum Showing thePresence of the Y-MPS on the Surface of the Glass 11

5. DRIFT Spectrum of Y-MPS/Amine Modified ParticulateGlass Showing the Increase in the Intensities of theY-MPS with the Percent Loading of the Treating Solution . . 12

6. Plot of the Integrated Intensity of the CarbonylBand of Y-MPS on the Surface of the ParticulateGlass as a Function of Weight Percent Loading ofthe Particulate Glass 12

7. Spectrum of Y-MPS on the Surface of the Glass in theCarbonyl Stretching Region Showing the VibrationalBand Assignments Including the C = 0 and C = C Modes .... 13

8. Plot of the Integrated Carbonyl Band as a Function ofWeight Percent Loading 14

9. Spectra of Y-MPS on Sunadex Glass Before and AfterHydrothermal Treatment Showing the Extensive Removalof the Silane with Water Exposure 15

10. Spectra of Y-MPS on the Surface of Glass Spheres WhichHave Been Acid-Washed to Prepare the Surface Prior toTreatment with the Silane 15

11. DRIFT Spectrum of Pure EVA 16

12. DRIFT Spectrum Showing the Silane at the EVA/GlassInterface 16

13. DRIFT Spectrum of Silane at the Interface BetweenGlass and EVA (Top Spectrum) and the Spectrum of theY-MPS Silane of the Glass 17

14. DRIFT Spectrum of the EVA/Y-MPS-Peroxide Before andAfter Hydrothermal Cycling 26

15. FTIR Spectra of Primed EVA/Glass Bead Specimens,Hydro thermally Aged at 60 °C 26

Vlll

Figures (Cont'd)

16. Infrared Spectra of Y-MPS Films with Three DifferentThicknesses Deposited on Aluminum Mirrors 30

L7. Infrared Spectra of: (A) Thick, and (B) Thin(Monolayer) Films of Y-MPS on Aluminum Mirrors 30

18. Ellipsometry Parameter A Versus Immersion Time inWater at 40°C for an Uncoated Aluminum Mirror 32

19. Ellipsometry Parameter A Versus Immersion Time inWater at 40°C for EVA Cured with Lupersol 101Laminated on Vapor-Deposited Aluminum 32

20. Ellipsometry Parameter A Versus Immersion Time inWater at 40°C for EVA Cured with TBEC Laminated onVapor-Deposited Aluminum 33

21. Scanning Electron Micrograph of Back Side of"Type 1" Silicon Cell (5000X) 34

22. EDAX: (A) Silicon and (B) Al Maps of Back Side of SameSilicon Cell as in Figure 21 35

23. Auger Electron Spectra of Back Side of Same SiliconCell as in Figure 21 36

24. Scanning Electron Micrograph of Back Side of "Type 2"Silicon Cell (2000X) 36

25. X-ray Photoelectron Spectrum of Back Side ofAs-Received "Type 2" Silicon Cell 37

26. XPS Spectrum of Laminated "Type 2" Silicon Cell AfterPeeling EVA Film 38

27. XPS Spectrum of Laminated "Type 2" Silicon Cell AfterBoiling 1 h and Peeling EVA Film 38

28. Reflection Infrared Spectra Obtained from AluminizedBack Side of "Type 1" Silicon Cell 40

29. Reflection Infrared Spectrum Obtained from AluminizedBack Side of "Type 1" Silicon Cell, Boiled in Waterfor 30 min, and Then Coated with a Thin Film of EVA .... 40

30. Reflection Infrared Spectra Obtained from AluminizedBack Side of a "Type 1" Silicon Cell Coated with a ThinFilm of EVA 41

IX

Figures (Cont'd)

31. Reflection Infrared Spectra Obtained from the Back Sideof a "Type 1" Silicon Cell Coated with a Thin Filmof EVA, But No Primer 41

32. Reflection Infrared Spectra Obtained from the Back Sideof a "Type 1" Silicon Cell Coated with Thin Filmsof A-11861 Primer and EVA 42

Tables

1. Adhesive Bonding Test Formulations 9

2. Hydrothermal Aging Conditions 9

3. Hydrothermal Aging of Glass Bead-Filled PolymerSpecimens at 40°C 19



6. Effect of Drying Out the Hydrothermally AgedSpecimens Reported in Table 3 22



x

SECTION I

INTRODUCTION

Adhesives and primers are necessary for the high reliability bonding ofphotovoltaic (PV) module interfaces to ensure structural integrity and longlife. The adhesion between the pottant and other components (i.e., substrate,superstrate, and outer cover) must have adequate bond strength to hold themodule components together, with a Flat-Plate Solar Array (FSA) Project goalof 30 years of outdoor service life. Accordingly, the material components ofthe module should be held together by strong chemical bonding, with theinterfacial bonding being resistant to delamination from the combinedweathering actions of moisture, temperature, and ultraviolet light.

Toward this end, Dow Corning (Dr. Edwin Plueddemann) is developingchemical bonding agents based on organosilane chemistry called silane couplingagents, which are chemically bifunctional. Using glass and ethylene vinylacetate (EVA) as an example, one of the chemical groups on a silane couplingagent is designed to react with glass, while the other chemical group willreact with EVA. Thus, the interfacial bonding is accomplished by actualprimary chemical bonds bridging across the interface.

In general, the silanes employed for these coupling agents are typicallyresistant to deterioration by ultraviolet light, and therefore the mainconcern is from hydrothermal deterioration over a long exposure timeoutdoors. Ordinarily the quality (strength) of bonded materials is monitoredby mechanical peel tests, where failure is defined as cohesive if it occurs ineither of the two bonded materials, or adhesive if failure occurs in theinterfacial region. Using silane coupling agents and, again, EVA and glass asan example, failures of control and hydrothermally aged EVA/glass specimensare normally cohesive within the EVA (References 1 and 2). That is, the bondstrength exceeds the tensile strength of EVA. Thus, accelerated hydrothermalaging may be decreasing the bond strength, but if it continues to exceed theEVA's tensile strength, it cannot be monitored by mechanical testing.

Dr. Koenig at Case Western Reserve University has been pioneeringtechniques based on Fourier Transform Infrared (FTIR) spectroscopy toinvestigate chemically bonded interfaces for information related to bondpopulation. The first part of this Project document is a report onDr. Koenig's FTIR activities at Case Western, in collaboration withDr. Miller. During this work, it was found necessary to disperse smalldiameter glass beads in EVA in order to increase the surface-to-volume ratioof test specimens for FTIR detection. Of technical significance, it wasrecently found (Reference 2) that the equilibrium water absorption of thesecomposite specimens was a strong function of the level of chemical bonding.For example, one series of unprimed EVA/glass specimens reported inReference 2 absorbed nearly 2000 wt % water at 60°C, whereas a primedEVA/glass specimen absorbed only 35 wt % water at 60°C. For future work, thisfinding offers a possibility for monitoring changes in the chemically bondedinterface from simple measurements of equilibrium-absorbed water contents ofaged specimens.

Dow Corning has successfully developed a silane coupling agentformulation (called a "primer") for EVA and glass. This primer is designatedA-11861, and experimental quantities are available from SpringbornLaboratories. Separately, work continues at Dow Corning to develop primersfor bonding EVA to solar cells, and solar cell metallization materials, suchas aluminum. The reason for this activity is that silane coupling agents alsoappear to passivate metal surfaces against corrosion (References 3, 4, 5,and 6).

Dr. Boerio at the University of Cincinnati has been pioneeringtechniques based on ellipsometry and reflection absorption infrared (RAIR)spectroscopy to investigate chemically bonded interfaces between metals andpolymers. The second part of this Project document is a report on activitiesat the University of Cincinnati. Of significance is the apparent finding thatthe silane coupling agent employed in the A-11861 EVA/glass primer may beeffective for bonding EVA to the aluminized back surfaces of solar cells, andmay also be effective in stopping environmentally induced corrosion of thisaluminum. Thus, a self-priming EVA formulated with this silane coupling agentwould work with both glass and solar cells. Dr. Boerio's work indicates thatthe silane coupling agent in A-11861 only works between EVA and the aluminumon solar cells, and not between EVA and any other aluminum. Althoughpreliminary, new insights into the general area of bonding to aluminum mayresult from this work.

SECTION II

CHEMICAL BONDING BETWEEN GLASS AND ETHYLENE VINYL ACETATE (EVA)

In order to meet the required specifications, encapsulated PV modulesmust remain intact for 30 years, reliably resisting delamination and separationof any of the component materials. For such long outdoor life, strong weather-stable interfacial bonds must be formed between the component materials. Inorder to achieve such stability, adhesion promoters are used to enhance thestrength and durability of such bonds. The chemical bonding technology hasbeen described in two previous reports (References 7 and 8) and should bereviewed prior to this report in order to ascertain the current status withrespect to materials, adhesion promoters, and process considerations.

Characterizing the chemical nature of the adhesion between differentmaterials is challenging and difficult. The first problem is the smallconcentration of interfacial bonds relative to the bulk. The second problemis that the interfacial bonds are generally quite similar to those of thebulk, making the detection more difficult. Efforts have been made tounderstand the mechanism of the coupling agent action. The earliest effortsfocused primarily on the role of the coupling agent in improving themechanical properties. Many potential coupling agents were tested and theresults generated a number of explanations ranging from improvements insurface wettability, presence of acid-base reactions, and the formation ofchemical bonds. Each of these postulates describe some aspect of the responseof the mechanical properties to the different silanes and their processes, butnone of them are sufficient to explain all of the observations.

Over the past 40 years, many techniques have been used to evaluate theinterfacial chemistry of materials. Some of these techniques includemicroscopy, contact angle measurements, gel permeation chromatography, zetapotentials, heats of immersion, and radioactive labeling. Although thesetechniques are precise in their analysis, the major limitation is that eitherthey do not yield molecular structural information or have sufficientsensitivity to detect monolayers. Infrared spectroscopy has been widely usedas a technique to study surfaces (References 9 and 10) and with the advent ofFTIR, it has become an even more effective tool because of the highsensitivity, excellent structural resolution, simple sampling techniques, andcomputer processing of the spectra. The effectiveness of FTIR for the studyof the interfaces of glass fiber composites has been demonstrated(References 11, 12, 13, and 14). On this basis, it was selected for use tostudy the interfacial bonding of glass/EVA in terrestrial PV modules.

There are a number of factors that determine the chemical basis of theadhesion. The nature of the surface functionality of the bonding materialsdetermines the potential for chemical modification of the adhesion. Anecessary prerequisite for molecular design of the adhesive interfacial bondsis knowledge of the surface functionalities. Using FTIR spectral reflectiontechniques, insights into the chemical nature of the surface can be obtained.

Adhesion is most often promoted using a coupling agent that is specificfor the materials to be bonded. A coupling agent, as the name implies, is amolecule that couples together two different molecules. This molecularfastening is made for the purposes of promoting adhesion between the twomolecules which are generally found at an interface between two surfaces. Acoupling agent is bifunctional so that each end of the molecule can connectwith the molecules of the surfaces that are being bonded. A especially usefulclass of coupling agents is the organosilanes, where one end of the moleculehas organic functionality to react with organic surfaces, and the other end isinorganic to react with inorganic surfaces.

Silane coupling agents have the general form X-(CH2)3-Si-(OR)3. Thefunctional group X can be varied to optimize the reactivity with the polymermatrix to be connected to the glass (Figure 1). The organic ethers (OR)groups are short chain ethers that are readily hydrolyzed in water that isslightly acidic to alcohols leaving a silanetriol group which is very reactivewith itself (condensation/polymerization) and the glass surface. Inindustrial practice, the silanes are dissolved in water or alcohol and theglass is treated with a dilute (1% by weight) solution. This dilute solutionis commonly referred to as a "primer." The amount of silane on the surface ofthe glass is very small and its molecular structure is a function of thetreating and drying conditions of the industrial process.

Beginning in the early 1960s, questions were being asked concerning thenature of the coupling agent on the surfaces. Is the coupling agent bound tothe substrate material? If so, is the type of bonding important to theadhesion? How much coupling agent is needed to yield maximum adhesion? Doesthe coupling agent chemically react with the polymer matrix? Is the amount ofchemical reaction important to the adhesion? Could the performance of thecoupling agent be controlled by the treating conditions? These questions canbe answered using FTIR in the appropriate manner.

The questions of lifetime of the adhesive bond are also complex. Is thecoupling agent present on the substrate material after exposure towater/humidity? Is the structure of the coupling agent modified by thepresence of water? Does the water preferentially attack the interfacialbonds? Questions of this kind can be answered with accelerated exposure testscombined with FTIR spectroscopic characterization.

The model of the adhesive joint for EVA/glass can be considered as asandwich structure (Figure 2) with a series of layers. These layers include:

(1) The bulk glass substrate.

(2) An interface between the glass and the coupling agent.

(3) A region of bulk polysiloxane (interphase).

(4) An interface between the coupling agent and the EVA.

(5) The bulk EVA region.

It is convenient to discuss the results of this report in terms of thesecomponents as each plays a role in the ultimate properties of the adhesion.

ORIGINAL PAGE ISOF POOR QUALITY

GENERAL STRUCTURE

X

TRIALKOXY Rn c, ORn\j ^^ oi — unSILANE |

OR

PHYSICALABSORPTION

X

1HO— Si —OH

10

H101

X

H20 I— HO ^i

iOH

OH

SOLVENT EVAPORATIONAND HYDROGEN

X

' s H *•HO -S i— O

1 '•' H '0:^HH|O1

BONDING

X

1' 0-Si— OH

1o.\H*

H10

1

X IS POLYMER REACTIVE GROUPSUCH AS

• METHACRYLATE• AMINE• EPOXY• STYRENE

SURFACEPOLYMERIZATION

X

* 1Si — 0 -

10

1

X X

1 1Si - O —Si — 0

1 10 0

I 1

/£ Si - 0 - Si -///y/////////////

- S i - O - S i - O - S i - O - - Si - O - Si - 0 - Si - 0 -

Figure 1. Bonding of Silane Coupling Agents to Glass

CHEMICALLY BONDEDINTERFACE

AINTERPHASE

A.

OPEN CIRCLES INDICATE REGIONS OF COUPLING AGENT,FILLED CIRCLES INDICATE REGIONS OF EVA

Figure 2. Interdiffusion Model for a Silane-Primed EVA/Glass Joint

A. FOURIER TRANSFORM INFRARED (FTIR) SPECTROSCOPY OF SURFACES AND INTERFACES

FTIR provides great potential for surface studies. Infrared spectroscopyis one of the most versatile, widely used techniques for the determination ofmolecular structure. Recent advances in infrared spectroscopy also offer newsampling methods to characterize surfaces of materials and absorbed species onthe surfaces. These methods include reflection and emission techniques. Thesenew sampling techniques are further enhanced using FTIR, rather than the conven-tional dispersive-type spectrometer. Moreover, data processing techniques suchas spectral subtraction and computer-scale expansion allow improved selectivity.

In FTIR, a Michelson interferometer is used instead of the prisms andslits used in conventional infrared spectroscopy. Obviously, a slitlessspectrometer has an advantage in energy throughput. However, a more pronouncedadvantage comes from the so-called Fellget's advantage, or multiplex advantage,which allows all resolution elements to be detected simultaneously throughoutthe spectral range; a dispersive instrument receives a signal only from oneresolution element at a time. The signal-to-noise ratio is proportional tothe square root of the number of resolution elements (i.e., the spectral rangedivided by the resolution). Each scan takes only 1 s, so the signal-to-noiseratio can be further increased by consecutive accumulation of scans as thescan rate is on the order of 1 s.

In FTIR, an interferogram is obtained that consists of many waves ofdifferent frequencies. This interferogram is collected as a function ofretardation time of the moving mirror that separates the light frequenciesaccording to their time of travel which is a function of the wavelength. Thisinterferogram is the Fourier transform of the desired infrared spectrum. Thespectrum is digitized so that it can be stored on the dedicated minicomputerfor further data processing.

1. Advantages of FTIR for Surface Studies

Surface studies are always a challenge to the limitations ofinstruments and methods. Surface studies require sufficient sensitivity, ahigh signal-to-noise ratio, and selectivity. It is apparent that the amountof material at the surface (four or five atomic layers) is negligible comparedto the bulk. In a transmission experiment with a pressed disc, the bulkequivalent of 103 monolayers must be penetrated. Therefore, the suppressionof the signal of the bulk (i.e., selectivity) is an essential factor.Furthermore, the dynamic range of the instruments and techniques need to belarge, as is the case for FTIR.

2. Sampling Techniques for FTIR Spectroscopy Studies

The various techniques for the study of surfaces using FTIR havebeen reviewed recently (Reference 15). For this work, two techniques wereused: transmission and diffuse reflectance (DRIFT). A discussion of thetransmission technique is unnecessary, as the technique is routine. However,the digital subtraction approach is important and will be described in thenext section.

Although diffuse reflection sampling has been used in the ultravioletand visible region for a long time, it was only with the increased sensitivityof FTIR that diffuse reflectance was possible in the infrared region. Thebasic principle behind diffuse reflectance spectroscopy is that light incidentupon a solid or powdered surface is diffusely scattered in all directions(Reference 16). The scattered light is collected with the proper optical setupand directed to the infrared detector for analysis. Thus, diffuse reflectancepermits the study of surfaces of particulate and rough samples to be studied.The principal advantages of diffuse in the infrared are: (1) it is a surfacespectroscopy technique which is quantitative, and (2) it allows opaque and/orstrongly absorbing solids and other problem samples to be studied withrelative ease.

Diffuse reflectance rather than transmission methods were used in thisstudy for several reasons, the first and foremost being the increasedsensitivity of diffuse reflectance to the coupling agent on the glass surface.This sensitivity made the observation of very small amounts of coupling agentspossible. These measurements would have been difficult with ordinarytransmission techniques. Also, it is not necessary to alter the sample in anyway to obtain the spectrum. For transmission studies, the sample must beground to a small particle size, dispersed in potassium bromide,and pressed into a pellet. The grinding process may destroy those chemicalbonds of direct interest in adhesion. Normally, in diffuse reflectance, thepowdered sample is mixed with powdered KBr and placed in a sample chamber.

3. Data Processing Techniques

In spite of the development of the reflection methods indicatedabove, the study of surfaces using transmission measurements.remains the mostwidely used technique, particularly for inorganic and organic substrates whichcan be prepared in the appropriate form. It was recognized that absorbancesubtraction could be used for the study of surfaces. If the infrared spectrumof a bulk sample is obtained and the same bulk sample is given a surfacetreatment, spectral subtraction can be used to remove the interference. Thebulk phase and the resultant difference spectrum would then reflect thedifferences in the surfaces before and after treatment. If the untreatedsample is subtracted from the treated sample, positive absorbance bands wouldrise from the higher concentration of surface species in the treated sample,and negative bands from species would be preferentially found in the untreatedsurfaces. After the isolation of the difference spectrum, the absorbance scalecan be expanded to one-third of the signal-to-noise ratio, as the strongabsorbing bands due to the bulk have been removed, giving a very highsensitivity for the technique. In this manner, one can determine the natureof the chemical reactions occurring with surface treatments.

B. EXPERIMENTAL PROCEDURES (MATERIALS)

Sunadex glass is preferred for the solar modules, but it was anticipatedand confirmed that the flat glass would not have sufficient surface area toallow detection of the coupling agent by FTIR, so particulate glass was usedto increase the surface-to-volume ratio. The specimens were prepared bydispersing high loadings of extremely fine, essentially spherical glass

powder^ in the EVA 'and curing as usual to "set" the bond chemistry. The testspecimens were prepared to determine the effects of hydrothermal aging (heatand water) on glass/primer/EVA bonds. These specimens were designed for usein a parallel program in which the decay of mechanical properties (bondstrength) could be measured with simultaneous spectroscopic investigation ofthe interface to follow the chemistry.

The test specimens consisted of both primed and unprimed glass powdersand both EVA (Elvax 150) and polyethylene were used as the polymers. The glasspowders were first washed in dilute mineral acid to remove alkaline residueson the surface,' then dried and treated with primer^ until a total of 2 wt %of active primer was deposited on the surface. The treated glass was thenused to prepare the formulations described in Table 1.

Tensile bars were cut from cured molded plaques and then exposed tohydrothermal aging conditions by immersion in water at the times andtemperatures described in Table 2. The test bars resulting from theseexposures were then tested for mechanical properties before and after drying,and also for spectroscopic examination.

C. FTIR SPECTROSCOPY STUDIES OF UNAGED EVA/GLASS INTERFACES

The FTIR spectra of each of the components in the EVA/glass-couplingagent has been isolated as indicated in Figure 3. Spectrum A is EVA. Thespectrum shown as B is obtained from the total system containing all of thecomponents. Spectra C are the spectra of the interfaces involving the silaneat the EVA/glass interface. Spectrum D is the Y-MPS coupling-agentinterphase, and Spectrum E is the particulate glass. The discussion thatfollows will explain the nature of these spectra and the information theycontain.

1. Glass

The diffuse reflectance spectrum (DRIFT) of the untreated orpristine glass is shown in Figure 4. The strong bands observed below1320 cm~l are the symmetric and asymmetric Si-O-Si stretching modes. Thespectrum shows little absorption in the organic alcohols (OH) stretchingregion indicating very few siloxyl groups on the surface.

1-Potter's Industries, Hasbrouck Heights, New Jersey, Product 5000. SphericalA-glass powder, 400 mesh, mean particle diameter approximately 20u, meansurface area, 0.269 m^/g.

^Springborn Laboratories, Primer A11861, consisting of 90% Dow Corning Z-6030,9% of Benzyl dimethylamine catalyst, and 1% Lupersol TBEC peroxide. Theseactive ingredients are dissolved in methyl alcohol to give a 1% solution.Equivalent to approximately 100 monolayers of silane. The Z-6030 couplingagent is chemically Y-methacryloxypropyltrimethoxy silane, commonly referredto as Y-MPS.

Table 1. Adhesive Bonding Test Formulations

18181-A 18181-B 18181-C 18181-D

Elvax 150 (EVA)

Chemplex 6230 polyethylene

Lupersol TBEC

Unprimed glass beads

Primed glass beads

Volume percent loading

Cure: 1508C/20 min

100 100

100 100

1.5 PHR 1.5 PHR 1.5 PHR 1.5 PHR

110.8 PHR — 110.8 PHR

110.8 PHR — 110.8 PHR

30% 30% 30% 30%

all

PHR = Parts per hundred parts EVA or polyethylene.

Table 2. Hydrothermal Aging Conditions

Water Temperature Hours

100 250 500 1000 2000 5000

40°C

60°C

80°C

X X X X

X X X X

X . X X X

X = Time and temperature conditions tested.

4000 2900 1800 700

(B) COMPOSITE

(EVA/TBEC/GLASS/MPS)

18152-B

I I I I I I I I I3800 3400 3000 2600 2200 1800 1400 1000 600

(C) COMPOSITE INTERFACE

(SILANE AT EVA/GLASS INTERFACE) (SILANE AT GLASS/EVA INTERFACE)

X 20 X 20

3000 2860 2820 2680I I I I I I I I I I

3020 2880 2840 2800 2660 18001780176017401720170016801660

(D) COUPLING AGENT INTERPHASE X 20

I J4000 3500 3000 2500 2000

WAVENUMBERS

1500 1000

Figure 3. Diffuse Reflectance Spectrum of MaterialComponents and Composites

10

MPS/AMINE MODIFIED PARTICULATE GLASS

PARTICULATE GLASS

DIGITAL SUBTRACTION (x 10)

4000 3500 3000 2500

WAVENUMBERS

2000 1500 1000

Figure 4. Diffuse Reflectance FTIR Spectra of Y-MPS/Amine Modified ParticulateGlass, the Untreated Particulate Glass, and the Difference SpectrumShowing the Presence of the Y-MPS on the Surface.of the Glass

2. Coupling Agent on Glass Surface

The spectrum of the treated glass is shown in Figure 3 (shownas C). The bottom spectrum (shown as E) shows the scale expanded (X10)difference obtained by subtracting the spectrum of the untreated glass fromthe treated glass. This difference reflects the spectrum of the couplingagent on the glass surface.

The amount of coupling agent on the surface is determined by the concen-tration of the coupling agent in the aqueous solution. The DRIFT spectra ofparticulate glass that has been treated with different concentrations of theY-MPS/amine agent are shown in Figure 5. Below a weight percent of 0.03, nocoupling agent is detected on the surface. The amount on the surface increasesas the concentration of treating solution increases. Figure 6 shows therelationship between the integrated intensity of the carbonyl band of the Y-MPSand the weight percent loading in the treating solution. A nearly linearcorrelation is found. Calculations indicate that at 1% loading, there areapproximately 50 monolayers of coupling agent on the surface. This result is inagreement with previous measurements of this kind for glass fibers (Reference 17),

Figure 7 shows the band assignments in this region with the 1720 cm~lband being assigned to the free carbonyl stretching mode and the 1700 cm~lto the H-bonded carbonyl mode. This latter mode is probably associated with

11

0 14 I I I

MPS/AMINE MODIFIED PARTICULATE GLASS

000|_1800 1750 1700 1650

WAVENUMBER, cm-1

1600 1550

Figure 5. DRIFT Spectrum of Y-MPS/Amine Modified Particulate Glass. Showing theIncrease in the Intensities of the Y-MPS with the Percent Loading ofthe Treating Solution

1200

100 0

> 800

c/)z

z0 60 0

orO

? 40 0

200

I 1 I

MPS/AMINE MODIFIED GLASS

00

^

1 12 3 4

WEIGHT PERCENT LOADING

Figure 6. Plot of the Integrated Intensity of the Carbonyl Band of Y-MPS onthe Surface of the Particulate Glass as a Function of Weight PercentLoading of the Particulate Glass

12

Z

cc

3 5

30

2 5

2 0

1 5

1 0

0 5 I I1800 1750 1700

WAVENUMBERS

1650 1600

Figure 7. Spectrum of Y-MPS on the Surface of the Glass in the CarbonylStretching Region Showing the Vibrational Band Assignments Includingthe C = 0 and C = C Modes. (The band at 1700 cnr1 is the carbonylstretching mode arising from hydrogen bonding of the initial layer ofY-MPS to the silanols on the glass surface.)

the hydrogen bonding of the Y-MPS to the glass surface and arises primarilyfrom the first monolayer of Y-MPS on the surface of the glass. The weakerband at 1640 cm~l is assigned to the C = 0 stretching mode of the Y-MPS. Thisband (at 1640 cm""-'-) is especially useful for following the polymerization ofthe Y-MPS to polymethacrylate.

3. Hydrothermal Stability of the Glass/Coupling Agent Interface

In an effort to determine the hydrothermal stability of thecoupling agent on the glass surface, the treated glass powder was exposed towater at 80°C for 3 h, and the spectra measured afterward. The amount ofcoupling agent remaining on the surface is indicative of the relativehydrolytic stability of the coupling agent on the surface (Reference 18).Figure 8 shows the results of these measurements by comparing the amount ofcoupling agent on the surface before and after water exposure. The resultsindicate that with the amine catalyst, very little of the coupling agent isremoved by the water exposure, although without the amine (except for theinitial monolayer on the surface), all of the coupling agent is removed. Thebehavior observed for the coupling agent on glass fibers without aminecatalyst indicates that most of the coupling agent can be easily removed(Reference 19). The amine catalyst is apparently effective in condensing andcrosslinking the coupling agent interphase on the surface making theinterphase hydrothermally stable.

13

Figure 8.

140.0

120 0

1000

1-CA>

w 800

Qtu

< 600OLU1-z

400

200

OOf

1 1 1 1 1

O BEFORE HYDROTHERMAL TREATMENT

— X AFTER HYDROTHERMAL TREATMENT X _

AFTER HYDROTHERMAL TREATMENTWITHOUT THE AMINE IN THE „TREATING SOLUTION U

XX

o ~o

— —

X

X0

x orxx-CP °-t — , , .- .-yoo^-^ ^ 1 1 r 1

Plot ofLoading

1 2 3 4 5 6

WEIGHT PERCENT LOADING

the Integrated Carbonyl Band as a Function of Weight Percent

However, it has been observed that the nature of the glass surface playsan important role in the hydrothermal stability. Figure 9 shows the spectrumbefore and after hydrothermal treatment of Sunadex Glass, and it is apparentthat nearly all of the coupling agent has been removed. This problem can bealleviated by an acid pretreatment of the glass as shown in Figure 10. In thelatter case, the only effect of the water is to hydrolyze the residual methoxygroups of the coupling agent.

A. EVA

The spectrum of EVA obtained by DRIFT is shown in Figure 11. Thebands near 2800 cm~l are the carbon-hydrogen stretching modes of thepolyethylene backbone. The sharp band near 1800 cm~l is associated with thecarbonyl group of the vinyl acetate copolymer, and the remaining bands arehydrocarbon in nature.

5. EVA/Glass Interface

In Figure 12, the spectrum of an EVA/glass composite is shown aswell as the spectrum of the silane modified glass. The final spectrum is thesubtracted spectrum which isolates the spectrum of the silane at the EVA/glassinterface. It is apparent that the silane carbonyl mode can be observed, butthe overlap of the acetate carbonyl band is extensive, making quantitativemeasurements difficult. In the carbon-hydrogen stretching region, it can be

0 35

HYDROTHERMAL TREATMENT

0 003200 3100 3000 2900

WAVENUMBER , cm-1

2800 2700

Figure 9. Spectra of Y-MPS on Sunadex Glass Before and After HydrothermalTreatment Showing the Extensive Removal of the Silane with WaterExposure

040

0 30

> 0 20

m111H

£ 010

0 00

-0 10

50 LAYERS

HYDROTHERMAL TREATMENT

I I I I3200 3100 3000 2900 2800 2700

WAVENUMBER, cm -1

Figure 10. Spectra of Y-MPS on the Surface of Glass Spheres Which Have BeenAcid-Washed to Prepare the Surface Prior to Treatment with theSilane. (The spectra are shown before and after hydrothermaltreatment showing that the acid improves the stability.)

15

50000

40000

>- 30 000

- 20 000

10000

0 000 ±3800 3300 2800 2300

WAVENUMBER, cm-1

1800 1300

Figure 11. DRIFT Spectrum of Pure EVA

1 4

1 2

1 0

0 8

0 6

0 4

0 2

00 —

-0 2

C = O STRETCHING

EVA/GLASS COMPOSITE

SILANE MODIFIED GLASS

SILANE AT EVA/GLASS INTERFACE

I I I I1800 1780 1760 1740 1720 1700 1680 1660

WAVENUMBER,

Figure 12. DRIFT Spectrum Showing the Silane at the EVA/Glass Interface. (Thetop spectrum is the EVA/glass spectrum, the middle spectrum is thespectrum of the Y-MPS modified glass, and the bottom is thedifference spectrum obtained by subtracting the treated glassspectrum from the composite. The bands observed are associatedwith the interfacial silane structure.)

16

observed that structural changes occur as a result of the peroxide-inducedpolymerization of the coupling agent. This polymerization is indicated by theloss of the carbon-hydrogen stretching mode of the vinyl group as indicated inFigure 13. Quantitative determinations of the amount of Y-MPS polymerizationcan be made. The calculations indicate that nearly 80% of the Y-MPS haspolymerized with the peroxide, and only 2Q% of the Y-MPS has polymerized inthe absence of the peroxide. This polymerization of the Y-MPS produces aninterlocking interpenetrating network with the EVA (see Figure 2). Thisinterpenetrating network has nonhydrolyzable carbon-carbon bonds as well as thehydrolyzable Si-O-Si bonds. The presence of the carbon-carbon bonds shouldcontribute to the hydrothermal stability of the EVA/glass interface.

D. HYDROTHERMAL AGING OF EVA/GLASS INTERFACE

Accelerated aging tests were performed by placing the model composites inwater at 40, 60, and 80°C for lengths of time up to 5000 h at present. Sampleswere removed at specified times and tested mechanically according to specificationASTM D-638, and also tested by FTIR spectroscopy.

1. Mechanical Testing

For mechanical testing, the following properties were measured:

(1) Modulus (1% strain).

(2) Ultimate tensile strength.

8 0

7 0

6 0

5 0

4 0

3 0

2 0

1 0

00

SILANE ATEVA/GLASSINTERFACE

SILANE ONSUBSTRATESURFACE

= CH2

3020 3000 2980 2960 2940 2920

WAVENUMBER, cm-1

2900 2880 2860

Figure 13. DRIFT Spectrum of Silane at the Interface Between Glass and EVA (TopSpectrum) and the Spectrum of the Y-MPS Silane of the Glass. (Thetop spectrum shows that the Y-MPS has polymerized during theformation of the interface because of the initiation by the peroxide.)

17

(3) Ultimate elongation.

(4) Dimensional change (length).

(5) Weight change (water absorption).

These measurements are tabulated for each temperature in Tables 3 through5, and Tables 6 through 8 give the results for the test specimens dried at 105°Cfor a period of 72 h. The purpose for this was to determine the extent to whichthe bond strength was reversible by simply removing the absorbed water.

At 40°C water immersion (see Table 3), little effect was noticed for allthe formulations except for 18181-A (EVA containing unprimed glass beads).This specimen slowly increased in water absorption from +0.15% weight gainafter 100 h, to +51% weight gain after 2000 h. Apart from a minor decrease inelongation, the properties of this compound remained unchanged.

The 60°C water immersions (see Table 4) were found to be more severe,and then both the compounds with and without primer began to respond. The EVAwithout primer gained an enormous +2000% increase in weight at the 2000-h testpoint. The mechanical properties also shifted dramatically at this point andthe compound lost 90% of its modulus, 96% of its tensile strength, and about90% of its elongation. The equivalent compound with primed glass gained only+35% in weight; most of the other properties were only slightly affected. Theequivalent polyethylene specimens gained only 3 to 4% water and wereunaffected by the immersion. In the case of EVA, the difference between primedand unprimed glass filler is noticeable and suggests that the properties atthe interface (adhesion bond) are responsible for the observed effects.

The next immersion temperature was 80°C (see Table 5) and resulted inrapid deterioration of the EVA specimens containing unprimed glass. After500 h, the specimen had increased 500% in weight, and lost 80% of its tensilestrength and elongation. Because of softening and tearing, this specimen didnot survive the 1000-h immersion test point and was removed from furtherexposure. The EVA formulated with primer beads gained 61% by weight of waterat the 1000-h test point and retained most of its other properties reasonablywell. As with the previous experiments, the polyethylene compounds loadedwith primed and unprimed glass beads showed only small gains from waterabsorption and were virtually unchanged.

It appears that the surface of the glass or the interface between the glassand the polymer is extremely hygroscopic and is effective in trapping largeamounts of water in the polymer. The primer is of help in inhibiting thisphenomenon and seems to give at least a ten-fold reduction in the amount of waterabsorbed. Water absorptions in the bulk polymer without fillers has been deter-mined in a previous set of experiments (Reference 1). For EVA, the equilibriumwater content at 40 and 60°C is on the order of +0.3%. At 80°C, pure EVA (Elvax150) equilibrates at approximately 1% by weight water. These water absorptionexperiments indicate that the glass/polymer interface is the region affected.

A second set of test bars from the water immersion tests were dried todetermine the extent of property recovery and give an indication as to whenpermanent and non-reversible damage had occurred. The results of theseexperiments are given in Tables 6 through 8. For the 40°C immersions, all the

18

Table 3. Hydrothermal Aging of Glass Bead-Filled Polymer Specimens at 40°C

Specimen Property3 Control 100 h 500 h 2000 h

18181-A,EVA, WithoutPrimer

18181-B,EVA, WithPrimer

18181-C,Polyethylene,WithoutPrimer

18181-D,Polyethylene,With Primer

MODUT

UF, %Awt %AL %

MODUT

UF, %Awt %AL %

MODUT

UF, %Awt %AL %

MODUT

UF, %Awt %AL %

28301380600

——

2500905350——

8 x 10*2180~0—

—

1.6 x 105

434020

——

1.6 x 103

1295600

+0.15%0

1.6 x 103

1070385

+0.19%0

1.2 x 105

199510

+0.52%0

1.7 x 105

423510

+0.31%0

1.7 x 103

1230570

+18.9%+5.5%

1.8 x 103

900190+2%0

1.2 x 105

-22000

+1.4%0

1.6 x 105

-4380000

8.3 x 102

1210510

+51%+15.3

2.4 x 103

1150390

+3.5%+1.3

1.2 x 105

23600

+2.5%0

1.5 x 105

44900

+0.6%0

aMechanical properties by ASTM D-638, 10 in./minimum rate of strain,Awt % = percent weight gain; AL % = percent length gain-tensile bar.

19


Specimen





Property3

MODUT

UF, %Awt %AL %

MODUT

UF, %Awt %AL %

MODUT

UF, %Awt %AL %

MODUT

UF, %Awt %AL %

Control

28301380600—

—

2500905350

——

8 x 104

2,180~0

——

1.6 x 10'5

434020

——

100 h

1.8 x 103

1240570

+0.92%0

2 x 103

935445

+0.36%0

1.1 x 105

22305

+0.6%0

1.6 x 105

440010

+0.33%0

250 h

1.6 x 103

960515

+29.3%+10.8%

2.5 x 103

930285+4%0

1.4 x 105

2510~0

+1.4%0

1.5 x 105

4450~0

+0.7%0

500 h

3.8 x 102

530300

+410%+78.4%

2 x 103

990315

+6.3%+6%

1.1 x 105

2070~0

+2.5%0

1.2 x 105

4270~000

2000 h

3 x 101

5060

+2015%+179%

1.6 x 103

830120

+34.7%+12.5%

3 x 104

2710~0

+4.8%Ob

1.3 x 105

2980~0

+3.08%0


"Blistered appearance.

20


Specimen





Property3

MODUT

UF, %Awt %AL %

MODUT

UF, %Awt %AL %

MODUT

UF, %Awt %AL %

MODUT

UF, %Awt %AL %

Control

28301380600

——

2500905350

——

8 x 104

2180~0

——

1.6 x 105

434020

——

100 h

1.7 x 103

460365

+17.4%+26.71

2.2 x 103

987275+1%0

1.2 x 105

214510

+1.4%0

1.8 x 105

444510

+0.43%0

250 h

9.3 x 102

285120

+568%+109%

2.3 x 103

1010245

+12.9%+2.7%

1.3 x 105

2160~0

+3 . 5%+0.2%

1.7 x 105

3990~0

+0.7%0

500 h

1.1 x 103

310100

+503%+96 . 2%

3 x 103

910220

+16.9%+5.6%

1.3 x 105

2110~0

+4.1%+0.4%

1.4 x 105

3610~0

+2.9%+2.2%

1000 h

Removed*5

1.7 x 103

725140

+61.7%+21.5%

1.2 x 105

1930~0

+4.3%+0.6%

1.2 x 105

2910~0

+4.5%+0.2%


"Specimens too soft to be tested, break on handling.

21

Table 6. Effect of Drying Outa the Hydrothermally Aged SpecimensReported in Table 3

Specimen





Property^

MODUT

UF, %Awt %AL %

MODUT

UF, %Awt %AL %

MODUT

UF, %Awt %AL %

MODUT

UF, %Awt %AL %

Control

2.8 x 103

1380600

——

2.5 x 103

905350

——

8 x 104

21800

——

1.6 x 105

434020

——

100 h

2.4 x 103

1205590

——

2.2 x 103

965325

——

1.8 x 105

2790-0

——

1.8 x 105

4720~0

——

500 h

3.4 x 103

1280565

——

3 x 103

990315—

—

1.6 x 105

1770,~0

——

1.6 x 105

4571~0

——

2000 h

2.2 x 103

1210570

——

2.6 x 103

860260

——

1.5 x 105

3710~0

——

1.7 x 105

4380~0

——

aDried in circulating air oven, 105°C, 72 h.

^Mechanical properties by ASTM D-638, 10 in./minimum rate of strain,Awt % = percent weight gain; AL % = percent length gain-tensile bar.

22

Table 7. Effect of Drying Out3 the Hydrothermally Aged SpecimensReported in Table 4

Specimen





Property^

MODUT

UF, %Awt %AL %

MODUT

UF, %Awt %AL %

MODUT

UF, %Awt %AL %

MODUT

UF, %Awt %AL %

Control

2.8 x 103

1380600—

—

2.5 x 103

905350

——

8 x 104

21800

——

1.6 x 105

434020

——

100 h

2.5 x 103

1270590

——

2.6 x 103

955290

——

1.6 x 105

2580~0

——

1.9 x 105

4530~0——

250 h

3 x 103

1340575

——

3.1 x 103

890240——

1.7 x 105

3680~0——

1.8 x 105

4890~0——

500 h

2.7 x 103

960500

——

3.1 x 103

980240——

1.8 x 105

3050~0——

1.4 x 105

4510~0——

2000 h

1.1 x 103

480280——

2.8 x 103

850240——

1.8 x 105

26800

——

1.6 x 105

35900

——


^Mechanical properties by ASTM D-638, 10 in./minimum rate of strain,Awt % = percent weight gain; AL % = percent length gain-tensile bar.

23

Table 8. Effect of Drying Out3 the Hydrothermally Aged SpecimensReported in Table 5

Specimen





Property^*

MODUT

UF, %Awt %AL %

MODUT

UF, %Awt %AL %

MODUT

UF, %Awt %AL %

MODUT

UF, %Awt %AL %

Control

2.8 x 103

1380600

——

2.5 x 103

905350——

8 x 104

21800

——

1.6 x 105

434020——

100 h

2.7 x 103

1175545

——

2.6 x 103

905240

——

1.4 x 105

3160~0

——

1.6 x 105

4415~0

——

250 h

1.8 x 103

560290

——

2.6 x 103

920310

——

1.8 x 105

3280~0——

1.5 x 105

4100~0

——

500 h 2000 h

(see footnote c)

2.8 x 103

1020405—

—

1.7 x 105

2440~0

——

1.6 x 105

3800~0

——

2.6 x 103

880260

——

1.8 x 105

3010~0——

1.8 x 105

3710~0

——


^Mechanical properties by ASTM Dr638, 10 in./minimum rate of strain,Awt % = percent weight gain; AL % = percent length gain-tensile bar.

cBroken, removed from further testing.

24

specimens recovered almost all their original properties. At 60°C, signs ofpermanent damage now became apparent in the EVA specimen filled with unprimedglass beads. At the 500-h point, the tensile strength decreased by 30%, andat the 2000-h point, it had decreased to 65% of control. About a 50% reductionof elongation at break was also found at this point. The other EVA compoundcontaining primed glass recovered most of its control properties upon drying.The 80°C test specimens gave much the same performance except that now non-reversible effects seem to start at after about 250 h in the unprimed EVA/glassspecimen. Primed EVA/glass specimens survived 2000 h at this temperaturewithout damage.

2. Spectroscopic Testing

As reported in the previous section, the mechanical properties asa function of hydrothermal aging of the EVA composites without primer had sub-stantial declines, whereas when the glass surface is treated with the primersolution, the composites show a remarkable resistance to loss of mechanicalproperties, as a function of hydrothermal aging time and temperature.

An additional result of the accelerated hydrothermal aging is theincorporation of water into the composite structure, generally absorbing waterto many times their original weight.

For chemical information using FTIR spectroscopy, a preliminaryexperiment was carried out involving an accelerated hydrothermal aging of anEVA/glass composite. The test consisted of an 85°C water treatment for2 months with subsequent drying at 100°C in vacuum. The FTIR spectra of thesample over the region 2200 to 3800 cm"1 before and after the hydrothermaltreatment are shown in Figure 14. No major chemical differences were observedexcept for the apparent uptake of water which, was not removed in spite of therather substantial drying. This water is presumed to be on the surface of theglass powder as it is well known that the interface is the preferred site ofthe sorbed water. Examination of the spectra of the hydrothermally treatedsample in other spectral regions showed no observable chemical degradationafter 2 months. The primed EVA/glass interface has resisted hydrothermalattack up to at least 2 months (1440 h) of accelerated testing. Thisrepresents extraordinary stability.

The EVA/glass composites (which had undergone up to 2000 h ofhydrothermal aging in water at 60°C, and whose aging properties are shown inTable 4) were also examined by FTIR spectroscopy. The FTIR spectra of thesesamples after 100, 250, 500, and 2000 h are shown in Figure 15, over theregion 3500 to 4000 cm"1. Up to 500 h, FTIR spectroscopy of this samplefound no evidence for chemical changes and, as with the preliminary sample,detected only water uptake.

However, at 2000 h of hydrothermal aging, a new infrared band occurredat 3680 cm-1 [see Figure 15(D)]. With this finding, the FTIR spectra of theunprimed EVA/glass composites which had been hydrothermally aged at 60°C weretaken to serve as comparative controls. Again, after 2000 h of aging, this newband at 3680 cm"1 was observed, and it was tentatively concluded that thiswas not related to the primer. Examination of the mechanical property data inTable 4 for the primed EVA/glass specimen hints at changes (after 2000 h) that

25

EVA/MPS/PEROXIDE

BEFORE AND AFTER HYDROTHERMAL CYCLING

>-t

z

200

15 0

10.0

5 0

003800 3600 3400 3200 3000 2800 2600 2400 2200

WAVENUMBER, cm-1

Figure 14. DRIFT Spectrum of the EVA/Y-MPS-Peroxide Before and AfterHydrothermal Cycling. (The uptake of water is shown by theincrease in the absorption between 3610 and 3200 cm^-.)

4000 3900 3800 3700

WAVENUMBER, cm~1

3600 3500

Figure 15. FTIR Spectra of Primed EVA/Glass Bead Specimens, HydrothermallyAged at 60°C

26

seem to be outside the data patterns observed for lower aging times. It istempting to interpret this band as spectroscopic evidence of some irreversibleinterfacial event related to hydrothermal attack, although assignment of thisband to a given chemical reaction is not clear at this time.

27

SECTION III

CHEMICAL BONDING BETWEEN SOLAR CELLS AND EVA

Ethylene/vinyl acetate (EVA) has been developed for use as anencapsulation material for photovoltaic devices. However, there are twoimportant problems associated with the use of EVA in such applications:(1) the adhesion of EVA to the aluminized rear surface "of silicon cells ispoor, especially during exposure to moisture or high relative humidities atelevated temperatures; and (2) most tests (such as the- peel test) that areused to evaluate the adhesion are destructive in nature. Such tests providelittle information about the onset of delamination processes and about thelikely lifetime of a bonded assembly.

The initial phase of this research had two primary objectives: (1) toinvestigate the use of silane coupling agents as primers to enhance thehydrothermal stability of adhesive bonds between EVA and aluminum, and (2) todevelop non-destructive optical techniques for determining the stability ofpolymer/metal interfaces during exposure to moisture at elevated temperatures.Toward this end, qualitative tests were carried out investigating the adhesionof EVA to several types of aluminum substrates, which included vapor-depositedaluminum mirrors on glass, and the aluminized back surfaces of commercial solarcells. The results obtained are described below.

A. ALUMINUM MIRRORS

Adhesive bonding between EVA and most aluminum surfaces without primers areweak and are weakened even further by exposure to warm water.- The hydrothermalstability of the EVA/aluminum interface is improved by the use of silane primers.

For studies with aluminum mirrors, reflection-absorption infrared (RAIR)spectroscopy has been used to determine the nature of the interface betweenY-MPS and the oxidized surface of aluminum. RAIR is a technique for obtaininginfrared spectra of thin films on metal substrates by reflecting infraredradiation polarized parallel to the plane of incidence off the metals at large,almost grazing angles. RAIR spectra obtained- from Y-MPS films deposited onaluminum mirrors from dilute methanol solutions were always characterized byabsorption bands near 1720 and 1100 cm~l that were assigned to carbonyl andsiloxane groups (Figure 16), regardless of the concentration of the solutions.However, spectra obtained from Y-MPS films that were deposited from methanolsolutions, and then rinsed with methanol, were characterized by bands near 1700and 930 cm~l (Figure 17). The band near 1700 cm~l was assigned to carbonylgroups that were hydrogen bonded to surface hydroxyl groups. The band near 930cm~l was assigned to silanol groups that were also hydrogen bonded to thesubstrate. It seems that the first layer of Y-MPS molecules is absorbed ontoaluminum by hydrogen bonding through both the silanol and carbonyl groups. Thismay help to explain the effectiveness of Y-MPS as a primer for enhancing thehydrothermal stability of adhesive bonds between EVA and substrates such asaluminum. If the first layer of Y-MPS is absorbed through two functionalgroups, it may be more difficult to displace from the surface than if it wasabsorbed only through, for instance, the silanol groups.

29 PRECEDING PAGE BLANK NOT FILMED

I 1'

05%

I 0 125%

I I I I1800 1600 1400 1200 1000

WAVENUM8ER, cm-1

800

Figure 16. Infrared Spectra of Y-MPS Films with Three Different ThicknessesDeposited on Aluminum Mirrors

0 025%

1800 1600 1400 1200 1000

WAVENUMBER, cm-1

800

Figure 17. Infrared Spectra of: (A) Thick, and (B) Thin (Monolayer) Filmsof Y-MPS on Aluminum Mirrors

30

Another important result obtained during the first phase of this researchwith aluminum mirrors was the development of ellipsometry as a non-destructiveprobe for monitoring the polymer/metal interface during exposure to anaggressive environment such as warm water. This technique involves the use ofan optical ellipsometer to measure the relative amplitude reduction (called Y),and the relative phase shift (called A) for the parallel and perpendicularcomponents of linearly polarized light reflected from a polymer-coated metalsubstrate that is immersed in warm water.

Results obtained for an uncoated aluminum mirror that was evaporatedonto a glass slide and then immersed in water at 40°C indicated that Ainitially decreased by several degrees before increasing somewhat, and thendecreased again (Figure 18). The first decrease in A was interpreted as thedissolution of the native oxide, and the second decrease indicatedreprecipitation of the hydroxide known as pseudoboehmite on the aluminumsurface.

The conversion of the native oxide on aluminum to the hydrated oxidepseudoboehmite is generally considered to be the cause of debonding ofpolymers from aluminum in the presence of warm water (Reference 20). In orderto obtain stable bonds between a rubbery polymer such as EVA and aluminum, itis necessary to inhibit that corrosion. These preliminary tests with uncoatedaluminum mirrors demonstrated that the corrosion reaction could be followed byRAIR and also by ellipsometry. The next series of experiments were intendedto learn if these same techniques could monitor the corrosion reaction througha coating of EVA.

B. ALUMINUM MIRRORS WITH EVA COATINGS

When a thin film of EVA containing the cross-linking agent Lupersol 101was laminated onto an unprimed, vapor-deposited aluminum mirror and thenimmersed in warm water, ellipsometry results similar to those shown inFigure 19 were obtained. An induction period lasting approximately 10 h wasobserved during which A changed very little. Thereafter, A increased slowly,but steadily. When a thin film of A-11861 primer was applied to the aluminumsubstrate before the substrate was laminated with EVA, the ellipsometry resultswere similar, but the rate of increase of A as a function of time after theinduction period was considerably lower.

When a thin film of EVA containing the cross-linking agent TBEC waslaminated onto an aluminum film that had been evaporated onto a glass slide andthen immersed in warm water, the ellipsometry results were quite different.In that case, almost no induction period was observed. Instead, A began to risealmost immediately (Figure 20).

These results were interpreted as follows:

(1) When EVA is cross-linked with Lupersol 101, the reaction proceedsmostly through the vinyl acetate groups.

(2) The ethylene units are relatively unaffected and remain in rathercrystalline regions.

31

01OJTJ

I-_l

L1JQ

142

138

134

130

126

30 60 90 120

IMMERSION TIME, mm

150 180

Figure 18. Ellipsometry Parameter A Versus Immersion Time in Water at 40°C foran Uncoated Aluminum Mirror

208

204

200

19610 100

IMMERSION TIME, h1000

Figure 19. Ellipsometry Parameter A Versus Immersion Time in Water at 40°C forEVA Cured with Lupersol 101 Laminated on Vapor-Deposited Aluminum

32

—<

265

255

245

235

O

I10 100

IMMERSION TIME, h

1000

Figure 20. Ellipsometry Parameter A Versus Immersion Time in Water at 40°Cfor EVA Cured with TBEC Laminated on Vapor-Deposited Aluminum

(3) Water does not diffuse through these crystalline regions veryquickly and a substantial induction period is observed beforewater reaches the aluminum surface.

(4) Once water reaches the surface, dissolution of the native oxideand reprecipitation of pseudoboehmite begin.

(5) The arrival of water at the interface and the precipitation ofpseudoboehmite both cause the parameter A to rise.

When EVA is cross-linked with TBEC, cross-linking occurs through boththe ethylene and vinyl acetate groups. In this case, there are no crystallineethylene regions and the diffusion of water to the interface is fast. Theellipsometry parameter A rises quickly during exposure to water at elevatedtemperatures.

Independent of the choice of cross-linking agent, these experimentsnevertheless demonstrated that the moisture-induced corrosion of the surfacesof aluminum mirrors could be detected by ellipsometry through a coating ofcross-linked EVA. In addition, there was some evidence that Y-MPS slowed,but did not stop the corrosion of these aluminum mirrors.

C. ALUMINIZED SOLAR CELLS

Two different kinds of aluminized back surfaces have been identifiedusing scanning electron microscopy. Both kinds are quite rough, but there aresignificant differences in their morphologies. The first type of aluminized

33

back surface consists of ridges separated by channels (Figure 21). Analysisof such surfaces by energy dispersive x-ray analysis (EDAX) showed that theycontained a great deal of silicon as well as aluminum. Generally speaking,the ridges were rich in silicon and the channels were rich in aluminum(Figure 22). These solar cells with a back surface mixture of aluminum andsilicon were designated "type 1" solar cells.

Auger electron spectroscopy (AES) was also used to characterize the backsurfaces of "type 1" cells. Auger spectroscopy is much more surface sensitivethan EDAX, but the results were similar. Silicon Auger bands were observednear 78 and 92 eV (Figure 23). The band near 78 eV is characteristic ofsilicon in an oxide and the band near 92 eV is indicative of elemental silicon.The aluminum AES bands were observed near 53 and 64 eV. These bands arecharacteristic of aluminum in the oxidized and elemental states, respectively.

The second type of aluminized back surface solar cells, designated as"type 2," was characterized by large numbers of small, irregularly shapedfeatures (Figure 24). X-ray photoelectron spectra (XPS) of such surfaces werecharacterized by bands near 74, 124, 285, and 531 eV (Figure 25). The XPSbands near 74 and 124 eV were assigned to aluminum in an oxidized state, andthe XPS band near 531 eV was assigned to oxygen in an oxide, indicating thatthe back surface of "type 2" cells was mostly composed of an aluminum oxide.The XPS band near 531 eV was assigned to carbon in absorbed hydrocarbons.There was essentially no silicon on the back surface of "type 2" cells.

Figure 21. Scanning Electron Micrograph of Back Side ofCell (5000X)

'Type 1" Silicon

34

ORIGINAL PAGE ISOF POOR QUALITY

(A)

Figure 22.

(B)

EDAX: (A) Silicon and (B) Al Maps of Back Side of Same SiliconCell as in Figure 21

35

t/5 3LU

/A I

Figure 23.

10 20 30 40 50 60 70 80 90 100

KINETIC ENERGY, eV

Auger Electron Spectra of Back Side of Same Silicon Cell as inFigure 21

Figure 24. Scanning Electron Micrograph of Back Side of "Type 2" Silicon Cell(2000X)

36

Figure 25.

100009000 8000 7000 600.0 5000 4000 300.0 2000 1000 00

BINDING ENERGY, eV

X-ray Photoelectron Spectrum of Back Side of As-Received "Type 2"Silicon Cell

Several different investigations were carried out to determine thefailure mechanisms for adhesive bonds between EVA and the aluminized backsurfaces of silicon cells. In one experiment, an EVA film was laminated ontoan unprimed "type 2" cell and then peeled off. The adhesion of the film tothe cell was very good and it was very difficult to peel the film. XPSspectra, obtained from the cell after the EVA film had been peeled off, werecharacterized by a strong band near 285 eV that was assigned to carbon(Figure 26). However, there was very little evidence of aluminum, indicatingthat the failure of the bond was cohesive within the EVA. In a repeatexperiment, EVA was laminated onto the backside of an unprimed "type 2" celland the cell was then boiled in water for 1 h. Once again, it was verydifficult to peel the EVA from the cell, indicating that the adhesion of thefilm to the cell was still quite good. However, some evidence for aluminumwas observed in the XPS spectra obtained from the cells, showing that thelocus of failure between EVA and aluminum was closer to the interface afterboiling in water (Figure 27), presumably because of the formation ofpseudoboehmite.

The tests just described were destructive in nature and, as indicatedearlier, the program intent is to develop non-destructive tests. Therefore,the application of optical techniques (such as ellipsometry and reflection-absorption infrared spectroscopy) for the non-destructive characterization ofadhesive bonds between EVA, and the aluminized back surface of silicon cellshave been investigated.

It was determined, however, that ellipsometry was not suitable for useon the back surfaces of the cells. These surfaces were much rougher than thesurfaces of aluminum mirrors and, therefore, accurate ellipsometry data couldnot be obtained. As a result, we have concentrated on the use of infrared

37

l.O 9000 800.0 700.0 6000 500.0 4000 3000 2000 1000 00

BINDING ENERGY, eV

Figure 26. XPS Spectrum of Laminated "Type 2" Silicon Cell AfterPeeling EVA Film

COo_x

Figure 27.

i —

100009000 8000 7000 600.0 500.0400.0 3000 2000 1000 00

BINDING ENERGY, eV

XPS Spectrum of Laminated "Type 2" Silicon Cell After Boiling 1 hand Peeling EVA Film

38

spectroscopy for the non-destructive characterization of the interface betweenEVA and the backside of silicon cells. Trial experimentation indicated that"type 1" solar cells were a better choice for RAIR examination because of muchhigher specular infrared reflectivity.

D. REFLECTION ABSORPTION INFRARED SPECTROSCOPY STUDIES

As described earlier, the failure mechanism of adhesive bonds toaluminum mirrors usually involves the conversion of the native oxide to thehydrated oxide known as pseudoboehmite (Reference 20) and, initially, it wasdesired to determine if exposure of solar cells to water at elevatedtemperatures also resulted in the formation of pseudoboehmite on the backsurfaces of these cells. Infrared spectra was obtained from the backside of a"type 1" cell before and after the cell was boiled in water for 1 h. Theresults are shown in Figure 28. The spectrum obtained from the cell beforeboiling was essentially flat and featureless, but the spectrum obtained afterboiling was characterized by a strong band near 1080 cm~l that wasindicative of pseudoboehmite.

Next, it was desired to determine if the formation of pseudoboehmitebeneath a film of EVA could be detected by RAIR. Accordingly, the infraredspectrum was obtained of a "type 1" cell that was boiled in water for 30 minand then coated with a thin film of uncrosslinked EVA that was several hundredsof angstroms in thickness (Figure 29). Strong bands characteristic of EVAwere easily observed near 1740 and 1255 cm~^, but the band near 1080 cm~lthat is characteristic of pseudoboehmite was also observed. These resultsshowed that pseudoboehmite could be detected beneath EVA films on siliconcells.

Next, a thin film of uncrosslinked EVA was applied to a "type 1" celland the cell was then boiled in water. Occasionally, the cell was removedfrom the water and examined by infrared spectroscopy. The results obtainedare shown in Figure 30. Initially, only the bands characteristic of EVA wereobserved near 1740 and 1255 cm"1-. However, after immersion of the cell inboiling water for 20 min, some changes were observed in the infrared spectraand, after 50 min, the band characteristic of pseudoboehmite was clearlyobserved near 1080 cm"1-.

In another series of experiments, infrared spectroscopy was used todetermine the effects of a cross-linked coating of EVA with and without primerA-11861 on the aluminized back surfaces of silicon cells. For the first testwithout primer, a thin film of EVA was applied to a "type 1" cell by dippingthe cell in a 0.25% solution of EVA in chloroform. The film was then cross-linked using Lupersol 101 and the cell was immersed in boiling water. The cellwas removed from the water at intervals of 30 min and examined using infraredspectroscopy. Initially, only bands characteristic of EVA were observed, butafter 30 min, the band characteristic of pseudoboehmite was clearly observednear 1080 cm"1- (Figure 31). After 60 min, the band near 1080 cm~l wasquite strong, indicating that hydration of the aluminum oxide beneath the EVAfilm was very rapid. Cross-linking of the EVA alone did not stop theformation of the pseudoboehmite.

39

oc/5

(/)<

A8

BEFORE IMMERSION IN BOILING WATER FOR 30 MINAFTER IMMERSION IN BOILING WATER FOR 30 MIN

2000 1800I I I

1600 1400 1200

WAVENUMBER, cm-1

I1000

Figure 28. Reflection Infrared Spectra Obtained from Aluminized Back Side of"Type 1" Silicon Cell

05

oct-oc

\ \ I2000 1800 1600 1400 1200

WAVENUMBER, cm~1

I1000

Figure 29. Reflection Infrared Spectrum Obtained from Aluminized Back Side of"Type 1" Silicon Cell, Boiled in Water for 30 min, and Then Coatedwith a Thin Film of EVA

en

(TI-

A = BEFORE IMMERSION IN BOILING WATERAND AFTER IMMERSION FOR

B = 20 MINC = 50 MIN

I I I _L2000 1800 1600 1400 1200

WAVENUMBER, cm-1

1000

Figure 30. Reflection Infrared Spectra Obtained from Aluminized Back Side ofa "Type 1" Silicon Cell Coated with a Thin Film of EVA

O

<<r

A = BEFORE BOILING IN WATER FOR 30 MINB = AFTER BOILING IN WATER FOR 30 MIN

1800 1700 1600 1500 1400 1300 1200 1100 1000

WAVENUMBER, cm -1

Figure 31. Reflection Infrared Spectra Obtained from the Back Side of a"Type 1" Silicon Cell Coated with a Thin Film of EVA, But No Primer

Finally, the experiment just described was repeated, except that a thinfilm of primer A-11861 was applied to the backside of the cell before the EVAfilm was applied and cross-linked. The initial spectra of the cell weresimilar to those of the unprimed cell, except that some bands characteristicof the primer were observed between 1200 and 1000 cm~l [Figure 32(A)].However, the band near 1080 cm~l was not observed even after the cell wasimmersed in boiling water for times as long as 90 min [Figure 32(B)], and itwas concluded that the primer inhibited the formation of pseudoboehmite on thealuminized back surfaces of "type 1" solar cells. This result is in dramaticcontrast to the behavior of aluminum mirrors, where EVA and the Y-MPS-basedprimer was not effective to stop pseudoboehmite formation. Althoughunexplained at this time, the possibilities of a self-priming EVA for bothglass and solar cells using a common primer is strongly indicated.

The results obtained thus far can be summarized as follows. The adhesionof EVA to the oxidized surfaces of smooth metals is marginal, but is improved bythe use of primers containing silane coupling agents such as Y-MPS. Ellipsometryis an excellent in-situ, non-destructive technique for monitoring the stabilityof interfaces between EVA and smooth, reflecting metals. The aluminized backsurfaces of silicon cells are very rough and porous. The adhesion of EVA to suchsurfaces is excellent, especially when primers containing Y-MPS are used.Failure of adhesive bonds between EVA and the aluminized back surface of siliconcells during exposure to moisture at elevated temperatures involves the trans-formation of the native oxide to the hydrated oxide known as pseudoboehmite.Ellipsometry cannot be used to follow the formation of pseudoboehmite on the backsurfaces of such cells because of their poor reflectivity. However, infraredspectroscopy can be used to determine the formation of pseudoboehmite under EVAfilms on the back surfaces of silicon cells.

A = BEFORE BOILING IN WATER FOR 90 MINB = AFTER BOILING IN WATER FOR 90 MIN

o(/5

5C/J

<

Figure 32.

1800 1700 1600 1500 1400 1300 1200 1100 1000

WAVENUMBER, cm"1

Reflection Infrared Spectra Obtained from the Back Side of a"Type 1" Silicon Cell Coated with Thin Films of A-11861 Primerand EVA

SECTION IV

REFERENCES

1. Willis, P.B., Investigation of Test Methods, Material Properties, andProcesses for Solar Cell Encapsulants, Eighth Annual Report, DOE/JPL-954527-84/27, Springborn Laboratories, Inc., Enfield, Connecticut,July 1984.

2. Willis, P.B., Investigation of Test Methods, Material Properties, andProcesses for Solar Cell Encapsulants, Ninth Annual Report, DOE/JPL-954527-85/28, Springborn Laboratories, Inc., Enfield, Connecticut,July 1985.

3. White, M.L., "Encapsulation of Integrated Circuits," Proceedings of theIEEE, Vol. 57, p. 1610, 1969.

4. Jaffe, D., "Encapsulation of Integrated Circuits Containing Beam LeadedDevices with a Silicone RTV Dispersion," IEEE Transactions on Parts,Hybrids, and Packaging, Vol. PHP 12, p. 182, 1976.

5. White, M.L., "Encapsulating Integrated Circuits," Bell LaboratoriesRecord, p. 80, March 1974.

6. Shar, N.L., and Feinstein, L.G., "Performance of New Copper-BasedMetallization Systems in an 85°C, 80% RH, Cl£ ContaminatedEnvironment," Proceedings of 1977 Electronic Components Conference,pp. 84-95, 1977.

7. Plueddemann, E.P., Chemical Bonding Technology for Terrestrial SolarCell Modules, JPL 5101-132, Jet Propulsion Laboratory, Pasadena,California, September 1, 1979.

8. Coulter, D.R., Cuddihy, E.F., and Plueddeman, E.P., Chemical BondingTechnology for Terrestrial Photovoltaic Modules, JPL Publication 83-86,Jet Propulsion Laboratory, Pasadena, California, November 15, 1983.

9. Hair, M.L., Infrared Spectroscopy in Surface Chemistry, Marcel Dekker,Inc., New York, 1967.

10. Little, L.H., Infrared Spectra of Absorbed Species, Academic Press,London, 1966.

11. Ishida, H., and Koenig, J.L., J. Coll. and Inter. Sci., Vol. 64, p. 555,1978.

12. Chiang, C.H., and Koenig, J.L., J. Polym. Sci., Polym. Phys. Ed.,Vol. 20, p. 2135, 1982.

13. Graf, R.T., Koenig, J.L., and Ishida, H., J. Adhesion. Vol. 16, p. 97,1983.

43

14. Graf, R.T., Koenig, J.L., and Ishida, H., Anal. Chem., Vol. 56, p. 773,1984.

15. Culler, S.R., Ishida, H., and Koenig, J.L., The Use of Infrared Methodsto Study Polymer Interfaces, Ann. Rev. Mater. Sci., Vol. 13, p. 363,1983.

16. Kortum, G., Reflectance Spectroscopy, Springer-Verlag, New York, 1969.

17. Graf, R.T., Koenig, J.L., Ishida, H., Anal. Chem., Vol. 56, p. 773, 1984.

18. Ishida, H., and Koenig, J.L., J. Polym. Sci., Polym. Phys. Ed., Vol. 18,p. 233, 1980.

19. Ishida, H., and Koenig, J.L., J. Polym. Sci., Polym. Phys. Ed., Vol. 18p. 1931, 1980.

20. Boerio, F.J., Gosselin, C.A., Dillingham, R.G., and Liu, H.W., Analysisof Thin Films on Metal Surfaces, J. Adhesion, Vol. 13, pp. 159-176, 1981.

44

TECHNICAL REPORT STANDARD TITLE PAGE

1. Report No. ^ 2. Government Accession No.

4. Title and Subtitle

Chemical Bonding Technology: DirectInvestigation of Interfacial Bonds

7. Authorfc)J . Koenig

9. Performing Organization Name or

JET PROPULSION LAB<California Institul4800 Oak Grove DriiPasadena, Calif orn.

id Address

DRATORYte of Technology/eLa 91109 .

12. Sponsoring Agency Nome and Address

NATIONAL AERONAUTICS AND SPACE ADMINISTRATIONWashington, D.C. 20546

3. Recipient's Catalog No.

5. Report DoteJanuary 1986

6. Performing Organization Code

8. Performing Organization Report No.

10. Work Unit No.

11. Contract or Grant No.NAS7-918

13. Type of Report and Period Covered

JPL Publication

14. Sponsoring Agency Code

15. Supplementary Notes Sponsored by the U.S. Department of Energy through InteragencyAgreement DE-AI01-85CE89008 with NASA; also identified as DOE/JPL 1012-120 and asJPL Project No. 5101-284 (RTOP or Customer Code 776-52-61).

16. Abstract

This is the third Flat-Plate Solar Array (FSA) Project document reportingon chemical bonding technology for terrestrial photovoltaic (PV) modules. Theimpetus for this work originated in the late 1970s when PV modules employingsiliconne encapsulation materials were undergoing delamination during outdoorexposure. At that time, manufacturers were not employing adhesion promotersand, hence, module interfaces in common with the siliconne materials were onlyin physical contact and therefore easily prone to separation if, for example,water were to penetrate to the interfaces. Delamination with siliconnematerials virtually vanished when adhestion promoters, recommended by siliconnemanufacturers, were used. This report describes the activities related to thedirect investigation of chemically bonded interfaces.

17. Key Words (Selected by AutHor(j))

Organic Chemistry

18. Distribution Statement

Unclassified-unlimited

19. Security Clossif. (of this report)Unclassified

20. Security Clossif. (of this poge)Unclassified

21. No. of Poges54

22. Price

JPL 01M N9IU

Documents

Chemical Bonding Technology: Direct Investigation of ... · of hydrocarbon encapsulants such as ethylene vinyl acetate (EVA), the need ... D. HYDROTHERMAL AGING OF EVA/GLASS INTERFACE